JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR
1997, 68, 193–217
NUMBER 2 (SEPTEMBER)
DRUG DISCRIMINATION UNDER A CONCURRENT
FIXED-INTERVAL FIXED-INTERVAL SCHEDULE
D. E. MCMILLAN, MI LI, AND W. C. HARDWICK
UNIVERSIT Y OF ARKANSAS FOR MEDICAL SCIENCES
Pigeons were trained to discriminate 5.0 mg/kg pentobarbital from saline under a concurrent ?xed-
interval (FI) FI schedule of food presentation on which, after pentobarbital administration, responses
on one key were reinforced with food under an FI 60-s component and responses on the other key
were reinforced under an FI 240-s component. After saline administration, the schedule contingen-
cies on the two keys were reversed. After both pentobarbital and saline, pigeons responded more
frequently on the key on which responses had been programmed to produce the reinforcer under
the FI 60 component of the concurrent schedule. The schedule was changed to concurrent FI 150
FI 150 s for drug-substitution tests. In each bird, increasing doses of pentobarbital, ethanol, and
chlordiazepoxide produced increases in the proportion of responses on the key on which responses
had been reinforced under the FI 60 component after pentobarbital administration during training
sessions. The proportion of responses on that key was slightly lower for ethanol than for chlordiaz-
epoxide and pentobarbital. At a dose of pentobarbital higher than the training dose, responding
decreased on the key that had been reinforced under the FI 60 component during training sessions.
Phencyclidine produced less responding on the key programmed under the FI 60-s component than
did pentobarbital. Methamphetamine produced responding primarily on the key on which responses
had been reinforced under the FI 60-s component after saline administration.
Key words: drug discrimination, concurrent ?xed-interval schedules, matching law, drugs, key peck,
pigeons
One of the most important determinants
may reset the contingency on the ‘‘injection-
of drug effects in behavioral pharmacology is
appropriate’’ key. Investigators usually have
the schedule of reinforcement that maintains
reinforced responding under identical FR
responding (Kelleher & Morse, 1964), but
schedule values after administration of both
the effects of the reinforcement schedule
drug and vehicle (Colpaert, 1987), although
have received little attention in drug-discrim-
?xed-inter val (FI) schedules (Krimmer,
ination experiments. Recent experiments
McGuire, & Barry, 1984; Massey et al., 1992),
from our laboratory have suggested that the
variable-inter val (VI) schedules (Gouvier,
schedule of reinforcement is also a powerful
Akins, & Trapold, 1984), tandem schedules
determinant of the effects of drugs in drug-
(Witkin, Carter, & Dykstra, 1980), multiple
discrimination experiments (Massey, McMil-
schedules (McMillan & Hardwick, 1996;
lan, & Wessinger, 1992; Snodgrass & McMil-
Snodgrass & McMillan, 1991), second-order
lan, 1991, 1996).
color-tracking schedules (McMillan, Cole-Ful-
The usual procedure in drug-discrimina-
lenwider, Hardwick, & Wenger, 1982), and
tion experiments is to reinforce responses un-
concurrent schedules (Snodgrass & McMil-
der a ?xed-ratio (FR) schedule of reinforce-
lan, 1996) have also been used. Usually these
ment on only one response key if a drug has
schedules have been symmetrical, in that re-
been administered before the session (drug
sponding on the drug key after drug admin-
key) and to reinforce responses under the
istration and responding on the saline key af-
same schedule on a different key if the drug
ter
vehicle
administration
have
been
vehicle has been administered before the ses-
reinforced using the same schedule values.
sion (saline key). Responses on the inappro-
priate key never produce the reinforcer and
However, it has been shown that responding
can be biased toward responding on either
These experiments were supported by NIDA Grant DA
the drug or the saline key by manipulation of
02251 to D. E. McMillan.
the schedule of reinforcement (Koek & Slan-
Reprint requests should be addressed to D. E. McMil-
gen, 1982; McMillan & Wenger, 1984).
lan, Department of Pharmacology and Toxicology, Slot
If the schedule of reinforcement is an im-
611, University of Arkansas for Medical Sciences, 4301
West Markham Street, Little Rock, Arkansas 72205
portant determinant of stimulus control in
(E-mail: mcmillandonalde@exchange.uams.edu).
drug-discrimination experiments, it is likely
193

---

194
D. E. MCMILLAN et al.
that the shape of the generalization curve
tinguishing ‘‘incorrect’’ responses under the
when other drugs are substituted for the
appropriate stimulus conditions. Thus, as
training drug will also depend on the sched-
Colpaert (1986) has noted, the discrimina-
ule of reinforcement during discrimination
tion is established under ‘‘all-or-none’’ rein-
training sessions. There has been consider-
forcement contingencies, which might be ex-
able discussion about the relationship be-
pected to favor nominal responding.
tween dose and response in drug-substitution
Although it is generally true in practice
experiments (Colpaert, 1986). Most investi-
that responding on the incorrect key is not
gators have assumed that there is a quantita-
reinforced, this is not a necessary condition
tive relationship between dose and response,
for establishing drug discrimination, because
so that the dependent variable (e.g., percent-
we have shown that drug discrimination can
age of responses on the drug key) is mea-
be established under conditions of relative re-
sured on a continuous scale. When respond-
inforcement (Snodgrass & McMillan, 1996).
ing on the drug key is measured on a
This was done by reinforcing responding un-
continuous scale, the proportion of responses
der concurrent VI VI schedules of reinforce-
on the drug key after a dose of drug is usually
ment, on which the values of the component
considered to be a measure of the degree to
VI schedules provided different frequencies
which that dose is similar to the training dose.
of reinforcer delivery. After administration of
However, there are also those who maintain
the training drug, responding on both keys
that the relationship between dose and re-
was reinforced. However, there was a differ-
sponse is a quantal unit that does not vary
ence in the schedule of reinforcer delivery
(Mathis & Emmett-Oglesby, 1990). According
for responding on the two keys, with the
to this viewpoint, the subject detects either
schedule associated with the drug-biased key
the presence or absence of the training dose
(key that was programmed at the higher re-
when a drug is given; thus, responding is
inforcement density after drug administra-
measured on a nominal scale. Another pos-
tion) providing a 4:1 ratio of reinforcers com-
sibility is that whether responding in drug-
pared to the schedule associated with the
discrimination experiments is nominal or
vehicle-biased key (key that was programmed
continuous depends on the schedule of re-
at a higher reinforcement density after vehi-
inforcement. Holloway and Gauvin (1989)
cle administration). After saline administra-
have suggested that schedules that bias sub-
tion, the reinforcement schedules pro-
jects to maximize reinforcement in choice sit-
grammed on the two keys were reversed.
uations by con?ning responses to one alter-
Thus, under the concurrent schedule, rein-
native (e.g., simple FR schedules) generate
forcement contingencies were relative in con-
nominal responding, whereas schedules that
trast to the all-or-none contingencies that are
maximize reinforcement when animals dis-
usually employed in drug-discrimination ex-
tribute their responses according to the
periments.
matching law (Herrnstein, 1970, 1974) favor
We were successful in establishing drug dis-
a graded distribution of responding across
crimination under concurrent VI VI sched-
both response alternatives (e.g., VI sched-
ules (Snodgrass & McMillan, 1996). In these
ules).
experiments, pigeons responded about 70%
These suggestions are consistent with the
of the time on the key on which responses
data from our comparison between the ef-
could produce the reinforcer more frequent-
fects of FI and FR schedules of reinforcement
ly, which is close to the proportion of re-
on the form of the generalization gradient
sponses that the matching law predicts would
(Massey et al., 1992; Snodgrass & McMillan,
occur on that key. The procedure has several
1991). We found that graded responding oc-
potential advantages. First, the use of the con-
curred when responding was maintained un-
current schedule provided an opportunity to
der FI schedules and nominal responding oc-
integrate drug-discrimination data with the
curred when responding was maintained
matching law (Herrnstein, 1970). According
under FR schedules using both morphine
to the matching law, under concurrent inter-
and pentobarbital as training drugs. However,
val schedules of reinforcement, response per-
typically a drug discrimination is established
centages approximately match reinforcement
by reinforcing ‘‘correct’’ responses and by ex-
percentages in two-choice procedures across

---

DRUG DISCRIMINATION AND FI SCHEDULES
195
a wide range of schedule values. The match-
dine (PCP), chlordiazepoxide, ethanol, and
ing law not only provides mathematical pre-
methamphetamine. These drugs were used
dictions in choice situations but also provides
because they each produced different levels
a theoretical framework for accounting for
of substitution for the pentobarbital stimulus,
choice behavior (Mazur, 1991). Second, the
and, excluding ethanol, the same drugs and
use of concurrent schedules also provided
doses had been used in our previous study of
the opportunity to determine whether doses
drug administration under concurrent VI VI
of drugs that are substituted for the training
schedules in these same birds (Snodgrass &
dose can produce effects outside the range of
McMillan, 1996).
responding that was established on the two
keys during training sessions. Under the usu-
al schedules of reinforcement that maintain
METHOD
drug-discrimination responding, animals re-
Subjects
spond almost entirely on the drug key after
drug and almost entirely on the vehicle key
Three male White Carneau pigeons (Pal-
after vehicle. There is no opportunity to de-
metto Pigeon Plant), P257, P259, and P260,
termine whether doses of the training drug
served as experimental subjects. These birds
that are higher than the training dose can
previously had been used for drug discrimi-
produce more extreme responding on the
nation under concurrent VI VI schedules
drug key than the training dose does, because
(Snodgrass & McMillan, 1996). They were in-
the animal is already responding 100% of the
dividually housed with free access to food and
time on the drug key. However, under con-
water in a temperature- and humidity-con-
current interval schedules, the distribution of
trolled room that was maintained under a
responses on the two keys depends on the
12-hr normal phase lighting cycle. The pi-
schedule values. This allows the investigator
geons had been maintained at 80% of their
to use the schedule of reinforcement to de-
free-feeding weights in the previous experi-
termine the proportion of responses that will
ments and continued to be maintained at
occur on the two keys and provides the op-
these 80% weights for the duration of the
portunity for a greater proportion of re-
current experiments.
sponses to occur on the drug-biased key and
Apparatus
a lower proportion of responses to occur on
the saline-biased key during drug-substitution
The experimental chamber was a Ger-
tests than occurs during training sessions.
brands Model G5610 pigeon test cage en-
In the present experiments, we continued
closed in a Gerbrands Model G7211 sound-
our study of relative reinforcement of re-
and light-attenuating cubicle. Two 28-V DC
sponse alternatives in drug-discrimination ex-
lights illuminated the experimental chamber
periments by using concurrent FI FI sched-
during the session, except during a food cycle
ules to maintain responding. Concurrent FI
when a light over the food hopper was illu-
FI schedules differ from the concurrent VI VI
minated. On the front panel of the cage
schedules used in our previous experiments
three Gerbrands response keys (Model
(Snodgrass & McMillan, 1996) in that the
G7311) were mounted 7 cm apart, 20 cm
time of reinforcer delivery under FI sched-
above the grid ?oor. The center key was not
ules has a temporal regularity that makes re-
used in these experiments and remained
inforcer delivery potentially more predictable
darkened at all times. When operative, the
than it is under VI schedules. In parallel with
left key was blue and the right key was yellow.
our previous research with these same birds
A force of 0.15 N or greater was required to
on drug discrimination under the concurrent
operate these keys. A food hopper (Ger-
VI 60-s VI 240-s schedules (Snodgrass & Mc-
brands) containing mixed grain was accessi-
Millan, 1996), the present study established
ble to the pigeon for 4 s when scheduled con-
drug discrimination under a concurrent FI
tingencies
were
met.
A
desktop
micro-
60-s FI 240-s schedule. Pentobarbital also
computer, located in a room adjacent to the
served as the training drug in both studies.
room containing the experimental chamber,
Other drugs used to determine the speci?city
controlled the schedule and recorded the
of the pentobarbital stimulus were phencycli-
data through a MED Associates interface.

---

196
D. E. MCMILLAN et al.
Procedure
designation). Which of the concurrent sched-
ules was in effect depended upon whether
The methods for training these pigeons
pentobarbital or saline had been adminis-
have been discussed in detail previously
tered before the session. Responding was
(Snodgrass & McMillan, 1996). Immediately
maintained under these concurrent sched-
following completion of the experiments us-
ules for the duration of the study, with the
ing concurrent VI VI schedules (Snodgrass &
exception of control and test sessions.
McMillan, 1996), the schedule was changed
To limit reinforcement of switching be-
to a concurrent FI 60-s FI 240-s schedule of
tween keys (Catania, 1966), a changeover de-
reinforcement. Pigeons were trained to dis-
lay (COD) of 3 s was imposed, such that a
criminate pentobarbital (5.0 mg/kg, i.m.)
response could not produce a reinforcer un-
from saline under the concurrent FI 60-s FI
less it occurred at least 3 s after the bird had
240-s schedule of reinforcement. Following
switched responding to a different key. Thus,
an i.m. injection of 5.0 mg/kg pentobarbital
an adventitious association between a change-
or saline, birds were placed into the test
over response and reinforcer delivery was
chamber and a 10-min presession followed,
minimized. Training sessions were conducted
during which all the lights in the chamber
6 days per week. During initial training, drug
were extinguished and the keys were inoper-
or saline was administered according to a
ative. After the presession time had elapsed,
double-alternation sequence, but later on, a
the houselights were illuminated, and the
single-alternation schedule of pentobarbital
schedule contingencies were initiated. The
and saline administration (i.e., DSDSDS. . .)
training sessions terminated after the ?rst re-
was employed. For discussion purposes, the
inforcer delivery that occurred after 60 min
key on which responses were reinforced un-
had elapsed.
der the FI 60 component after pentobarbital
Under the discrimination training proce-
will be referred to as the pentobarbital-biased
dure, both the left and right keys were trans-
key, and the key on which responses were re-
illuminated at session initiation, and an FI
inforced under the FI 60 component after sa-
schedule was operative on each key. Thus, in
line will be referred to as the saline-biased
each session, the pigeons were placed in a
key.
choice situation in which two response alter-
Test sessions began when the subjects
natives were available. Responses on each key
reached the training criteria: The pigeons
were reinforced by 4-s access to mixed grain
had to complete at least 51% of their re-
in the presence of each of the training stimuli
sponses on the key on which responses pro-
each session, but the programmed rate of re-
duced the reinforcer under the shorter FI
inforcement differed by a ratio of 4:1. For Pi-
component for 12 consecutive training ses-
geon P257, the concurrent FI FI schedule was
sions, six each of pentobarbital and saline.
programmed to allow the bird to earn four
The percentage of responding on the appro-
times as many reinforcers on the right key
priate key had to be within
1 standard de-
(yellow key) compared to the number of re-
viation of the mean of the previous six ses-
inforcers earned on the left key (blue key)
sions. These criteria had to be met for both
after administration of the training drug. Af-
pentobarbital and saline administration prior
ter vehicle administration, Pigeon P257 could
to the initiation of substitution testing with
earn four times as many reinforcers on the
other drug doses. These criteria were
left key as could be earned on the right key.
reached in all birds after 35 sessions of train-
For Pigeons P259 and P260, the reinforce-
ing. It should be noted that these birds al-
ment contingencies after administration of
ready had considerable experience discrimi-
the training drug and saline were the reverse
nating pentobarbital from saline under a
of those for P257, so that the schedule during
concurrent VI VI schedule (Snodgrass & Mc-
the training sessions was a concurrent FI 60-s
Millan, 1996), and it is presumed that it
FI 240-s or a concurrent FI 240-s FI 60-s
would have required many more sessions to
schedule of reinforcement (hereafter it
meet these criteria had we initiated these ex-
should be understood that the schedule val-
periments with new birds. Test sessions under
ues are in seconds and the abbreviation for
the concurrent FI 150 FI 150 schedule were
seconds will be dropped from the schedule
conducted on Tuesdays and Fridays, with

---

DRUG DISCRIMINATION AND FI SCHEDULES
197
training sessions under the concurrent FI 60
key on which the response was made, and,
FI 240 schedule continuing on other days of
thereafter, each time the pigeon switched
the week. If a bird failed to reach criterion
keys, time accumulated for the other sched-
performance in a training session (less than
ule component), and number of reinforcers
51% stimulus-appropriate responding), test
earned under each schedule component
sessions were postponed until these criteria
were recorded. From these data the percent-
had been met under both the pentobarbital
age of responding on the pentobarbital-bi-
and saline training conditions.
ased key could be determined. The pento-
The procedure used during generalization
barbital-biased key was de?ned as the key that
tests was identical to the procedure used dur-
was associated with the FI 60 component after
ing training sessions, except that a concur-
the administration of pentobarbital during
rent FI 150 FI 150 schedule of reinforcement
the training sessions. The saline-biased key
was in effect. This FI 150-s schedule value was
was de?ned as the key that was associated
chosen because it is intermediate between
with the FI 60 component after administra-
that of the 60-s FI and the 240-s FI values. The
tion of saline during the training sessions.
schedule also had the advantage that none of
The percentage of responding on the pen-
the pigeons had experienced this reinforce-
tobarbital-biased key and the percentage of
ment rate during training sessions, so the re-
time allotted to responding on the pentobar-
inforcement schedule was less likely to pro-
bital-biased key produced very similar data
vide additional cues as to which response
(see Appendix). If a pigeon earned less than
would be reinforced more frequently during
half of the possible number of programmed
the drug-substitution sessions. The pentobar-
reinforcers, the data obtained at that dose
bital and saline training doses were adminis-
were not included in the analysis of dose–re-
tered to the pigeons under the concurrent FI
sponse effects.
150 FI 150 schedule prior to the administra-
tion of each dose–response curve and after
Drugs
all of the dose–response data had been col-
Pentobarbital sodium at doses of 1, 3, 5.6,
lected. These sessions were designated as con-
10, and 13 mg/kg (?rst determination of
trol sessions and measured the effect of the
dose–response effects) or 3, 5, and 10 mg/kg
schedule change that was used during drug-
(redetermination of pentobarbital effects
substitution tests on the stability of the stim-
with collection of cumulative response rec-
ulus control of behavior. Drug-substitution
ords) (Sigma Chemical Co.), phencyclidine
tests were conducted in single test sessions on
hydrochloride (PCP) at doses of 0.1, 0.3,
different days for each dose for each pigeon.
0.56, and 1.0 mg/kg (National Institute on
The training, control, and generalization test
Drug Abuse), methamphetamine hydrochlo-
sessions were 40 min in duration, and data
ride at doses of 0.3, 1.0, 1.8, 3.0, and 5.6
were collected for the entire session.
mg/kg (Sigma Chemical Co.), chlordiaze-
Finally, after dose–effect curves had been
poxide hydrochloride at doses of 1.0, 3.0, 5.6,
determined for all of the other test drugs that
7.8, and 10 mg/kg (Hoffman-La Roche,
were substituted for pentobarbital, the pen-
Inc.), and ethanol at doses of 0.25, 0.5, 0.75,
tobarbital dose–response curve was redeter-
1.0, and 1.5 g/kg were studied. All drugs ex-
mined. During this redetermination, re-
cept ethanol were dissolved in 0.9% physio-
sponses on each key were recorded separately
logical saline to concentrations that allowed
by two cumulative recorders to permit a more
an injection volume of 1 ml/kg and were ad-
detailed analysis of the relationship between
ministered intramuscularly into a breast mus-
patterns of responding on the two keys.
cle. Physiological saline was also used for ve-
hicle control injections. Doses are expressed
Data Analysis
as salts, except for ethanol.
The number of CODs, the number of re-
Ethanol (100%) was diluted to a 10%
sponses and rate of responding under each
weight/volume solution with tap water. The
FI component, time spent responding under
10% ethanol solution or tap water, which was
each FI component (the ?rst response at the
used as the vehicle control, was administered
beginning of the session began accumulating
through a rubber tube that was passed down
time on the FI component associated with the
the esophagus into the proventriculus 15 min

---

198
D. E. MCMILLAN et al.
Table 1
Means from six pentobarbital (Pb) training sessions and six saline (S) training sessions and
the percentage of responding, reinforcers earned, and time allocated on the pentobarbital-
biased key (%Pb) prior to the initiation of the testing phase in individual pigeons. The group
means are also shown.
Responses
Reinforcers
Time
Pigeon
Pb
S
%Pb
Pb
S
%Pb
Pb
S
%Pb
Pentobarbital training sessions
P257
3,120
546
85
37
8
83
1,896
333
85
P259
3,243
1,149
74
34
9
78
1,504
717
68
P260
2,392
519
82
36
9
80
1,745
471
79
M
2,918
738
80
36
9
80
1,715
507
77
Saline training sessions
P257
1,001
2,104
32
9
33
21
508
1,498
24
P259
1,064
2,121
33
9
36
20
605
1,614
27
P260
350
833
28
7
36
16
513
1,713
23
M
805
1,686
31
8
35
19
542
1,608
25
prior to session initiation. Doses for ethanol
of reinforcers earned on the same key. Thus
are expressed as grams per kilogram.
the birds undermatched responding under
the saline training conditions. The number of
CODs was similar for all 3 birds and did not
RESULTS
differ for pentobarbital and saline training
Table 1 shows the mean number of re-
sessions (see Appendix).
sponses, the mean number of reinforcers
Figure 1 shows the dose–effect curves for
earned, and the mean time allocated to re-
percentage of responding on the pentobar-
sponding on the pentobarbital-biased and sa-
bital-biased key as a function of pentobarbital
line-biased keys for each pigeon over the last
dose for both the initial determination of the
six pentobarbital and the last six saline train-
pentobarbital dose–response curve and the
ing sessions prior to the initiation of drug-sub-
redetermination of the effects of pentobar-
stitution testing. During the pentobarbital
bital to provide cumulative response records,
training sessions, Pigeons P257 and P259
which was done at the end of the study. Dur-
emitted a slightly higher total number of re-
ing the ?rst determination of the pentobar-
sponses than during saline training sessions,
bital dose–response curve, all birds showed
whereas Pigeon P260 emitted considerably
increased responding on the drug-biased key
more responses during pentobarbital training
as the dose of pentobarbital increased from
sessions than during saline training sessions.
1.0 mg/kg to 5.6 mg/kg. For Pigeons P257
However, the number of reinforcers obtained
and P259, the peak of the dose–effect curve
by each pigeon under the two training con-
was reached at the 10 mg/kg dose, whereas
ditions was nearly identical. Thus, the train-
for Pigeon P260, the peak occurred at a dose
ing dose of pentobarbital did not suppress re-
of 5.6 mg/kg. Perhaps of greater importance,
sponding, nor did it decrease the number of
at higher doses the curve descended for all
reinforcers earned, compared to saline. After
birds, but especially for P260. During the sec-
pentobarbital administration, the percent-
ond determination of the pentobarbital
ages of responses, time allocated to respond-
dose–response curve, for which a more lim-
ing, and reinforcers obtained on the pento-
ited range of doses was studied, the 3.0
barbital-biased key under the concurrent FI
mg/kg dose generated less responding on
60 FI 240 were close to the 80% predicted by
the drug-biased key for all 3 birds. The ten-
the matching law. After saline administration,
dency for pentobarbital to be less potent in
however, the percentages of responding and
producing responding on the drug-biased
time allocation on the pentobarbital-biased
key was also seen at the 5.0 and 10.0 mg/kg
key were slightly higher than the expected
doses for Pigeon P259 but not for the other
values (20%) compared with the percentage
2 birds. The 1.0 and 13.0 mg/kg doses of

---

DRUG DISCRIMINATION AND FI SCHEDULES
199
Fig. 1.
The dose–response curve for the effects of pentobarbital on the percentage of responses on the key on
which responses had been reinforced under the FI 60 component of the concurrent FI 60 FI 240 schedule after
pentobarbital administration during training for Pigeons P257, P259, and P260 and the group mean. Abscissae: dose
of pentobarbital (1, 3, 5, 5.6, 10, and 13 mg/kg). Ordinates: percentage of responses on pentobarbital-biased key.
Brackets at T show
1 standard deviation around the mean based on the data obtained during training sessions.
Brackets at C show
1 standard deviation around the mean based on the control sessions in which the schedule was
changed to concurrent FI 150 FI 150, which was the schedule used during determination of the dose–response
curves. The ?lled circles show the original pentobarbital dose–response curve, and the open circles show the rede-
termination of the effects of selected doses at the end of the study. The ?lled triangles and squares above T and C
show the effects of 5.0 mg/kg pentobarbital used during training and control sessions. The open triangles and squares
above T and C show the effects of saline injections during training and control sessions.
pentobarbital were not studied during the
the pentobarbital-biased key. For all pigeons,
second determination of the pentobarbital
as the number of responses emitted on the
dose–response curve.
pentobarbital-biased key increased or de-
Detailed data showing CODs, responses on
creased, the amount of time allocated to re-
each key, reinforcers earned on each key, and
sponding on that key also increased or
time allotted to responding on each key are
decreased. The dose–effect curves for per-
shown for pentobarbital and the other drugs
centage of time allotted to responding on the
for each bird in the Appendix. There were
pentobarbital-biased key peaked and then de-
no consistent effects on CODs with increasing
scended after higher pentobarbital doses, just
doses of pentobarbital, although occasional
as occurred for the dose–effect curves for
decreases in CODs were observed for Pigeon
percentage of responses on the pentobarbi-
P257 and increases in CODs for Pigeon P260.
tal-biased key.
Figure 2 shows the percentage of time
Figure 3 shows responses on the pentobar-
spent responding on the pentobarbital-biased
bital-biased key as a function of chlordiaze-
key as a function of pentobarbital dose. The
poxide dose. Chlordiazepoxide caused a
percentages of time allocated to the pento-
dose-dependent increase in the percentage of
barbital-biased key resulted in dose–effect
responses on the pentobarbital-biased key for
curves that were very similar to those based
Pigeons P257 and P259, whereas the dose–
on the percentages of responses emitted on
effect curve for P260 was more irregular. The

---

200
D. E. MCMILLAN et al.
Fig. 2.
The dose–response curve for the effects of pentobarbital on the percentage of session time allocated to
responding on the key on which responses had been reinforced on the FI 60 component of the concurrent FI 60 FI
240 schedule after pentobarbital administration during training sessions. Abscissae: dose of pentobarbital (1, 3, 5,
5.6, 10, and 13 mg/kg). Ordinates: percentage of session time allocated to responding on the pentobarbital-biased
key. Other details are as in Figure 1.
percentage of responding on the pentobar-
of responses on the pentobarbital-biased key
bital-biased key reached the levels attained
are shown in Figure 4. For Pigeon P257, with
under the pentobarbital training and control
increasing doses there was a graded increase
sessions. For P257, the peak in pentobarbital-
in the percentage of responses on the pen-
biased key responding occurred at the dose
tobarbital-biased key, and after the 1.5 g/kg
of 7.8 mg/kg chlordiazepoxide, and this bird
dose of ethanol, P257 responded on the pen-
failed to respond at the dose of 10.0 mg/kg;
tobarbital-biased key at the same percentages
for P259, asymptotic responding occurred on
as occurred during pentobarbital training ses-
the drug-biased key at the doses from 5.6 to
sions. Pigeon P259 produced a similar etha-
10.0 mg/kg; for P260, the peak effect oc-
nol dose–response curve, except for a small
curred at the dose of 7.8 mg/kg, and there
dip in the curve at the 1.0 g/kg dose. For
is a suggestion that at 10.0 mg/kg chlordiaz-
Pigeon P260, there again was a suggestion
epoxide the dose–response curve began to
that the dose–effect curve had risen to a peak
descend. Data for time allotted to responding
and then began to descend after the highest
on the drug-biased key were similar, except
dose of ethanol, although the decreases in
that the dose–response curve for P259 also
the percentage of responses on the pento-
began to descend after the highest dose (see
barbital-biased key after the 1.0 and 1.5 g/kg
Appendix). For Pigeons P257 and P259,
doses were very small when compared to the
there were decreases in CODs at two or more
0.75 g/kg dose. The mean dose–response
dose levels (see Appendix).
curve for ethanol was very similar to that for
Ethanol dose–response data for percentage
pentobarbital, except that the ethanol curve

---

DRUG DISCRIMINATION AND FI SCHEDULES
201
Fig. 3.
The dose–response curve for the effects of chlordiazepoxide on the percentage of responses on the key
on which responses had been reinforced under the FI 60 component of the concurrent FI 60 FI 240 schedule after
pentobarbital administration during training sessions. Abscissae: dose of chlordiazepoxide (1.0, 3.0, 5.6, 7.8, and 10.0
mg/kg). Ordinates: percentage of responding on pentobarbital-biased key. Other details are as in Figure 1.
did not reach as high a peak nor did it begin
the PCP dose–response curves did not reach
to descend at the highest doses tested. Data
the same asymptote as the pentobarbital
for the time allocated to responding on the
dose–response curves, and the curves did not
drug-biased key were similar to these data on
descend after the higher dose levels. When
percentage of responses on the drug key (see
the time allotted to responding on the drug-
Appendix). CODs were decreased after sev-
biased key was measured, effects similar to
eral doses of ethanol for Pigeon P257 and af-
those described for responding on the drug-
ter the highest dose for Pigeon P259, but Pi-
biased key were obtained (see Appendix). Pi-
geon P260 showed increases in CODs after
geons P257 and P259 showed decreases in
the two highest dose levels.
CODs at the high doses of PCP (see Appen-
The PCP dose–response data are shown in
dix).
Figure 5. For all 3 pigeons, the percentages
Figure 6 shows the percentage of respond-
of responding on the pentobarbital-biased
ing on the pentobarbital-biased key during
key after the 0.1 mg/kg dose of PCP were
the determination of the methamphetamine
approximately equal to the percentages un-
dose–response curve. When doses of meth-
der the training and control conditions after
amphetamine were administered, all birds re-
saline administration. Higher doses generally
sponded on the pentobarbital-biased key to
produced increasing responding on the pen-
about the same extent as they did under the
tobarbital-biased key, but the percentages of
training and control conditions after saline
responding on the pentobarbital-biased key
administration, although at some doses slight-
never reached the levels attained during pen-
ly more responding occurred on the pento-
tobarbital training and control sessions. Thus,
barbital-biased key than during saline train-

---

202
D. E. MCMILLAN et al.
Fig. 4.
The dose–response curve for the effects of ethanol on the percentage of responses on the key on which
responses had been reinforced under the FI 60 component of the concurrent FI 60 FI 240 after pentobarbital
administration during training sessions. Abscissae: dose of ethanol (0.25, 0.50, 0.75, 1.0, and 1.5 g/kg). Ordinates:
percentage of responding on pentobarbital-biased key. Brackets at W represent
1 standard deviation around the
tap water control mean (open circles), which is based on ?ve observations after tap water was administered down
the esophagus into the proventriculus. Other details are as in Figure 1.
ing sessions. Again, similar effects were
sponding is characterized by bursts of re-
observed for time allotted to responding on
sponding that look more typical of FR
the pentobarbital-biased key (see Appendix).
responding, followed by pauses. These bursts
Methamphetamine produced increases in
of responding are separated by periods of
CODs at low doses only in Pigeon P257.
pausing that are longer than those seen un-
Figure 7 shows the cumulative response
der the FI 60 component of the schedule.
records for all 3 birds during a training ses-
Close examination of the relationship be-
sion following saline and 5.0 mg/kg pento-
tween the pattern of responding under the
barbital. The cumulative records for the two
components of the concurrent schedule sug-
components of the concurrent schedule have
gests that many of the bursts of responding
been overlaid for comparison. For all 3 birds
under the FI 240 component occurred dur-
after saline administration a typical pattern of
ing the pauses under the FI 60 component.
responding appears to have developed under
If reinforcer deliveries under the FI 60 com-
the FI 60 component (Ferster & Skinner,
ponent are compared with reinforcer deliv-
1957, see p. 157), characterized by a pause
eries under the FI 240 component, the bursts
after the delivery of the reinforcer and end-
of responding after every fourth reinforcer
ing in a terminal rate of responding, which
delivery under the FI 60 component usually
may or may not have been preceded by a
result in a reinforcer delivery under the FI
brief period of positive acceleration in re-
240 component. These relationships between
sponding. Under the FI 240 component, re-
component schedules hold in the same reg-

---